For some reason, it was added to the Unix and VMS build templates, but
Windows was forgotten.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/22947)
Like in #17345, evp_extra_test links libcrypto statically, but also has
a dynamic/shared load via the legacy provider, which leads to ambiguous
behavior in evp_extra_test on some platforms, usually a crash (SIGSEGV)
on exit via the atexit handlers. Statically link the legacy provider to
avoid this.
Fixes#22819
Helped-by: Neil Horman <nhorman@openssl.org>
Helped-by: Tomas Mraz <tomas@openssl.org>
Signed-off-by: Randall S. Becker <randall.becker@nexbridge.ca>
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22905)
This function can be called during OPENSSL_cleanup() when
the cache was already flushed and deallocated.
Fixes#22939
Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22941)
On some systems it is too small although the system allows longer
filenames.
Fixes#22886
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Neil Horman <nhorman@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22889)
When a provider can't return parameters, make that a warning instead of an
error, and continue to list further providers.
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/22866)
We test its validity by trying to load it.
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/22866)
If the provider's output dispatch table is NULL, trying to parse it causes a
crash. Let's not do that.
Reviewed-by: Dmitry Belyavskiy <beldmit@gmail.com>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/22866)
Common symbols (type 'C' in the 'nm' output) are allowed to be defined more
than once. This makes test/recipes/01-test_symbol_presence.t reflect that.
Fixes#22869 (partially)
Fixes#22837
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22880)
On some platforms, the compiler may add symbols that aren't ours and that we
should ignore.
They are generally expected to start with a double underscore, and thereby
easy to detect.
Fixes#22869 (partially)
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22880)
HPE has a weird preference to prefix letters and zero-padding. Properly trim
them before processing.
CLA: trivial
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22891)
Do not allow mid-expression line breaks.
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22814)
Some platform implementations are without `sharedlib_import()`, so we need
to check that it exists before using it.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/22789)
$(SRCDIR)/util/libcrypto.num and $(SRCDIR)/util/libssl.num were made their
own targets to have 'make ordinals' reproduce them (run mknum.pl) only if
needed.
Unfortunately, because the shared library linker scripts depend on these
.num files, we suddenly have mknum.pl run at random times when building.
Furthermore, this created a diamond dependency, which disturbs parallell
building because multiple mknum.pl on the same file could run at the same
time.
This reverts commit 0e55c3ab8d.
Fixes#21999
Partially fixes#22841
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22890)
Coverity recorded issues 1551739 and 1551737, a potential double free in the
tests. It occurs when the DUP operation fails in such a way val3_read is
returned as the same pointer as val2_read. Ideally it should never
happen, but resetting val3_read to 0 should satisfy coverity that there
is no issue here
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/22800)
Coverity issue 1551719 noted CRYPTO_secure_used referenced a shared
variable without taking the appropriate read lock. Add that.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
(Merged from https://github.com/openssl/openssl/pull/22802)
Add null check to cmac_size(). This avoids a seg-fault encountered
with cmac when EVP_MAC_CTX_get_mac_size() is called before init.
Extend mac testing in evp_test.c to check that the sizes returned by
EVP_MAC_CTX_get_mac_size() before and after init make sense (this also
ensures that we no longer seg-fault).
Fixes#22842
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22858)
When the CMS_ReceiptRequest cannot be created,
the rct_to and rct_from may be leaked.
Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Todd Short <todd.short@me.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22742)
And clean up partially created choice objects, which have
still the default type = -1 from ASIdentifierChoice_new().
Fixes#22700
Reviewed-by: Todd Short <todd.short@me.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22745)
"d_i in RFC8017" -> "d_i" in RFC8017
Reviewed-by: Todd Short <todd.short@me.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22868)
well know -> well known
CLA: trivial
Reviewed-by: Todd Short <todd.short@me.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22778)
Signed-off-by: lan1120 <lanming@huawei.com>
Reviewed-by: Todd Short <todd.short@me.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22795)
The little-endian optimization is doing some type-punning in a way
violating the C standard aliasing rule by loading or storing through a
lvalue with type "unsigned int" but the memory location has effective
type "unsigned long" or "unsigned long long" (BN_ULONG). Convert these
accesses to use memcpy instead, as memcpy is defined as-is "accessing
through the lvalues with type char" and char is aliasing with all types.
GCC does a good job to optimize away the temporary copies introduced
with the change. Ideally copying to a temporary unsigned int array,
doing the calculation, and then copying back to `r_d` will make the code
look better, but unfortunately GCC would fail to optimize away this
temporary array then.
I've not touched the LE optimization in BN_nist_mod_224 because it's
guarded by BN_BITS2!=64, then BN_BITS2 must be 32 and BN_ULONG must be
unsigned int, thus there is no aliasing issue in BN_nist_mod_224.
Fixes#12247.
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22816)
`tls_parse_ctos_key_share()` didn't properly handle the option.
Avoid the need to deal with the option in multiple places by properly
handling it in `tls_parse_ctos_psk_kex_modes()`.
Reviewed-by: Todd Short <todd.short@me.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22844)
Added self-hosted runners for freebsd-x86_64 and ubuntu-aarch64.
Reviewed-by: Tom Cosgrove <tom.cosgrove@arm.com>
Reviewed-by: Anton Arapov <anton@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22804)
There were no PKIX-TA(0) SPKI(1) Full(0) (i.e. "0 1 0") test cases in
"danetest.in".
There is now at least a success case, which will exercise freeing the public
key after it is sanity checked, since with PKIX-TA(0) there's nothing we can do
with just the raw public key, a full chain to a local trust anchor is in any
case required.
The failure (to match) code path is already well oiled, but failure to decode
while adding malfored TLSA records could still use some additional tests...
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22821)
We don't need the decoded X.509 Full(0) certificate for the EE usages 1 and 3,
because the leaf certificate is always part of the presented chain, so the
certificate is only validated as well-formed, and then discarded, but the
TLSA record is of course still used after the validation step.
Added DANE test cases for: 3 0 0, 3 1 0, 1 0 0, and 1 1 0
Reported by Claus Assmann.
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22821)
Instead of trying to move the doomed sct back
to the src stack, which may fail as well, simply
free the sct object, as the src list will be
deleted anyway.
Reviewed-by: Paul Yang <kaishen.yy@antfin.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22762)
